This study investigates the influence of socioeconomic status, health literacy, and numeracy on treatment decisions and the occurrence of adverse events in patients with breast, lung, and prostate cancer within a Nordic healthcare setting.
A follow-up to a cross-sectional, mixed-methods, single-centre study.
A Nordic, tertiary cancer clinic.
A total of 244 participants with breast, lung and prostate cancer were initially identified, of which 138 first-line treatment participants were eligible for this study. First-line treatment participants (n=138) surpassed the expected cases (n=108).
Not applicable as this was an observational study.
The study’s primary endpoint was the rate of guideline adherence. The secondary endpoint involved assessing treatment toxicity in the form of adverse events.
Guideline-adherent treatment was observed in 114 (82.6%) cases. First-line treatment selection appeared uninfluenced by participants’ education, occupation, income or self-reported health literacy. A minority (3.6%) experienced difficulties following treatment instructions, primarily with oral cancer medications.
The findings indicated lesser cancer health disparities regarding guideline adherence and treatment toxicity within the Nordic healthcare framework. A causal connection may not be established; however, the findings contribute to discourse on equitable cancer health provision.
Personalised Exercise-Rehabilitation FOR people with Multiple long-term conditions (PERFORM) is a research programme that seeks to develop and evaluate a comprehensive exercise-based rehabilitation intervention designed for people with multimorbidity, the presence of multiple long-term conditions (MLTCs). This paper describes the protocol for a randomised trial to assess the feasibility and acceptability of the PERFORM intervention, study design and processes.
A multicentre, parallel two-group randomised trial with individual 2:1 allocation to the PERFORM exercise-based intervention plus usual care (intervention) or usual care alone (control). The primary outcome of this feasibility trial will be to assess whether prespecified progression criteria (recruitment, retention, intervention adherence) are met to progress to the full randomised trial. The trial will be conducted across three UK sites and 60 people with MLTCs, defined as two or more LTCs, with at least one having evidence of the beneficial effect of exercise. The PERFORM intervention comprises an 8-week (twice a week for 6 weeks and once a week for 2 weeks) supervised rehabilitation programme of personalised exercise training and self-management education delivered by trained healthcare professionals followed by two maintenance sessions. Trial participants will be recruited over a 4.5-month period, and outcomes assessed at baseline (prerandomisation) and 3 months postrandomisation and include health-related quality of life, psychological well-being, symptom burden, frailty, exercise capacity, physical activity, sleep, cognition and serious adverse events. A mixed-methods process evaluation will assess acceptability, feasibility and fidelity of intervention delivery and feasibility of trial processes. An economic evaluation will assess the feasibility of data collection and estimate the costs of the PERFORM intervention.
The trial has been given favourable opinion by the West Midlands, Edgbaston Research Ethics Service (Ref: 23/WM/0057). Participants will be asked to give full, written consent to take part by trained researchers. Findings will be disseminated via journals, presentations and targeted communications to clinicians, commissioners, service users and patients and the public.
2.0 (16 May 2023).
Preventing readmission to hospital after giving birth is a key priority, as rates have been rising along with associated costs. There are many contributing factors to readmission, and some are thought to be preventable. Nurse and midwife understaffing has been linked to deficits in care quality. This study explores the relationship between staffing levels and readmission rates in maternity settings.
We conducted a retrospective longitudinal study using routinely collected individual patient data in three maternity services in England from 2015 to 2020. Data on admissions, discharges and case-mix were extracted from hospital administration systems. Staffing and workload were calculated in Hours Per Patient day per shift in the first two 12-hour shifts of the index (birth) admission. Postpartum readmissions and staffing exposures for all birthing admissions were entered into a hierarchical multivariable logistic regression model to estimate the odds of readmission when staffing was below the mean level for the maternity service.
64 250 maternal admissions resulted in birth and 2903 mothers were readmitted within 30 days of discharge (4.5%). Absolute levels of staffing ranged between 2.3 and 4.1 individuals per midwife in the three services. Below average midwifery staffing was associated with higher rates of postpartum readmissions within 7 days of discharge (adjusted OR (aOR) 1.108, 95% CI 1.003 to 1.223). The effect was smaller and not statistically significant for readmissions within 30 days of discharge (aOR 1.080, 95% CI 0.994 to 1.174). Below average maternity assistant staffing was associated with lower rates of postpartum readmissions (7 days, aOR 0.957, 95% CI 0.867 to 1.057; 30 days aOR 0.965, 95% CI 0.887 to 1.049, both not statistically significant).
We found evidence that lower than expected midwifery staffing levels is associated with more postpartum readmissions. The nature of the relationship requires further investigation including examining potential mediating factors and reasons for readmission in maternity populations.
Annual cognitive screening in older adults is essential for early detection of cognitive impairment, yet primary care settings face time constraints that present barriers to routine screening. A remote cognitive screener completed on a patient’s personal smartphone before a visit has the potential to save primary care clinics time, encourage broader screening practices and increase early detection of cognitive decline. MyCog Mobile is a promising new remote smartphone-based cognitive screening app for primary care settings. We propose a combined construct and clinical validation study of MyCog Mobile.
We will recruit a total sample of 300 adult participants aged 65 years and older. A subsample of 200 healthy adult participants and a subsample of 100 adults with a cognitive impairment diagnosis (ie, dementia, mild cognitive impairment, cognitive deficits or other memory loss) will be recruited from the general population and specialty memory care centres, respectively. To evaluate the construct validity of MyCog Mobile, the healthy control sample will self-administer MyCog Mobile on study-provided smartphones and be administered a battery of gold-standard neuropsychological assessments. We will compare correlations between performance on MyCog Mobile and measures of similar and dissimilar constructs to evaluate convergent and discriminant validity. To assess clinical validity, participants in the clinical sample will self-administer MyCog Mobile on a smartphone and be administered a Mini-Cog screener and these data will be combined with the healthy control sample. We will then apply several supervised model types to determine the best predictors of cognitive impairment within the sample. Area under the receiver operating characteristic curve, accuracy, sensitivity and specificity will be the primary performance metrics for clinical validity.
The Institutional Review Board at Northwestern University (STU00214921) approved this study protocol. Results will be published in peer-reviewed journals and summaries provided to the study’s funders.
To assess the return on investment (ROI) of the New York Tobacco Control Programme (NY TCP).
New York and other states of the USA.
NY TCP.
Smoking prevalence, smoking-attributable healthcare expenditures (SAEs), smoking-attributable mortality, years of life lost (YLL), the dollar value of YLL and the ROI for healthcare expenditures and mortality.
We used a synthetic control method to estimate the effectiveness of NY TCP funding on smoking prevalence. The synthetic control method created a comparison group that best matched the adult smoking prevalence trend in New York state in the period prior to implementation of the NY TCP and compared smoking prevalence in the state to smoking prevalence in the synthetic control in the period after treatment (2001–2019). The synthetic control group represents what the trend in smoking prevalence in New York would have been had there been no tobacco control expenditures. The ROI was calculated as net savings for each outcome divided by net programme expenditures.
Cumulative savings in SAE in New York from 2001 to 2019 amounted to US$13.2 billion. An estimated 41 771 smoking-attributable deaths (SADs) were averted in New York from 2001 to 2019, and an estimated 672 141 YLL averted as a result of NY TCP funding in the same period. From 2001 to 2019, the ROI for SAE in New York was approximately 14, the economic value ROI of the YLL due to SAD was nearly 145 and the combined ROI was almost 160.
In this study, we found relatively large ROIs for the NY TCP, which suggests that the programme—which lowers SAE and saves lives—is an efficient use of public funds.
The aim of this scoping review was to provide an overview of current research into topical oxygen therapies including the under-researched singlet oxygen for wound healing. A scoping review was undertaken using five databases. After duplicates and ineligible studies were excluded, 49 studies were included for a narrative review. Out of the included 49 studies, 45 (91.8%) were published in the past 10 years (2013–2023) with 32 (65.3%) published in the past 5 years (2018–2023). Eight of the studies were systematic reviews and/or meta-analysis and 18 were RCTs. The search identified zero human RCTs on singlet oxygen, but one human cohort study and five studies in animals. There is evidence that topical oxygen therapy may be useful for the treatment of chronic wounds, mainly diabetic foot ulcers. Singlet oxygen has shown potential, but would need further confirmation in controlled human trials, including more research to understand the bio-properties.
Background and aims: Bacteria in wounds can lead to stagnation of wound healing as well as to local or even systemic wound infections up to potentially lethal sepsis. Consequently, the bacterial load should be reduced as part of wound treatment. Therefore, the efficacy of simple mechanical wound debridement should be investigated in terms of reducing bacterial colonisation. Patients and methods: Patients with acute or chronic wounds were assessed for bacterial colonisation with a fluorescence camera before and after mechanical wound debridement with sterile cotton pads. If bacterial colonisation persisted, a second, targeted wound debridement was performed. Results: A total of 151 patients, 68 (45.0%) men and 83 (55.0%) women were included in this study. The male mean age was 71.0 years and the female 65.1 years. By establishing a new analysis method for the image files, we could document that the bacterial colonised areas were distributed 21.9% on the wound surfaces, 60.5% on the wound edges (up to 0.5 cm) and 17.6% on the wound surroundings (up to 1.5 cm). One mechanical debridement achieved a significant reduction of bacterial colonised areas by an average of 29.6% in the wounds, 18.9% in the wound edges and 11.8% in the wound surroundings and was increased by performing it a second time. Conclusions: It has been shown that even a simple mechanical debridement with cotton pads can significantly reduce bacterial colonisation without relevant side effects. In particular, the wound edges were the areas that were often most contaminated with bacteria and should be included in the debridement with special attention. Since bacteria remain in wounds after mechanical debridement, it cannot replace antimicrobial therapy strategies, but offer a complementary strategy to improve wound care. Thus, it could be shown that simple mechanical debridement is effective in reducing bacterial load and should be integrated into a therapeutic approach to wounds whenever appropriate.
To explore researchers’ experiences of funding processes, the effort and burden involved in applying for funding, obtaining funding and/or fulfilling reporting requirements with a UK health and social care research funder.
A cross-sectional online survey study with open (free-text) and closed questions (August to November 2021).
Researchers with experience of applying for/obtaining funding and/or experience of fulfilling reporting requirements for UK health and social care research funded between January 2018 and June 2021.
The survey was completed by 182 researchers, of which 176 had experience with applying for/obtaining funding, and 143 had experience with fulfilling reporting requirements during the timeframe. The majority of the 176 respondents (58%) completed between 7 and 13 key processes in order to submit an application and 69% felt that it was critically important to undertake these key processes. Respondents (n=143) reported submitting an average of 17 reports as part of research monitoring to a range of organisations (eg, funders, Higher Education Institutions). However, only 33% of respondents felt it was critically important to provide the requested reporting information to the different organisations. Thematic analysis of free-text questions on application and reporting identified themes relating to process inefficiencies including streamlining and alignment of systems, lack of understanding of processes including a need for improved communication and feedback from organisations with clear explanations about what information is needed, when and why, the support required by respondents and the time, effort and impact on workload and well-being.
Through this study, we were able to identify funding processes that are considered by some to be effortful, but necessary, as well as those that were perceived as unnecessary, complex and repetitive, and may waste some researchers time and effort and impact on well-being. Possible solutions to increase efficiency and enhance value in these processes were identified.
by Anna Kis, Katinka Tóth, Lívia Langner, József Topál
When pre-treated with social stimuli prior to testing, dogs are more susceptible to human influence in a food preference task. This means, after a positive social interaction they are more willing to choose the smaller amount of food indicated by the human, as opposed to their baseline preference for the bigger amount. In the current study we investigate if and how various forms of social interaction modulate choices in the same social susceptibility task, testing dogs with varying early life history (pet dogs, therapy dogs, former shelter dogs). In line with previous studies, dogs in general were found to be susceptible to human influence as reflected in the reduced number of “bigger” choices in the human influence, compared to baseline, trials. This was true not only for pet dogs with a normal life history, but also for dogs adopted from a shelter. Therapy dogs, however, did not uniformly change their preference for the bigger quantity of food in the human influence trials; they only did so if prior to testing they had been pre-treated with social stimuli by their owner (but not by a stranger). Pet dogs were also more influenced after pre-treatment with social stimuli by their owner compared to ignoring and separation; however after pre-treatment by a stranger their behaviour did not differ from ignoring and separation. Former shelter dogs on the other hand were equally influenced regardless of pre-treatment by owner versus stranger. In summary these results show that dogs’ social susceptibility is modulated by both interactions immediately preceding the test as well as by long term social experiences.by Daniela Tamas, Nina Brkic Jovanovic, Stanka Stojkov, Danijela Cvijanović, Bozana Meinhardt–Injac
ObjectiveMost previous studies have examined emotion recognition in autism spectrum condition (ASC) without intellectual disability (ID). However, ASC and ID co-occur to a high degree. The main aims of the study were to examine emotion recognition in individuals with ASC and co-occurring intellectual disability (ASC-ID) as compared to individuals with ID alone, and to investigate the relationship between emotion recognition and social functioning.
MethodsThe sample consisted of 30 adult participants with ASC-ID and a comparison group of 29 participants with ID. Emotion recognition was assessed by the facial emotions test, while. social functioning was assessed by the social responsiveness scale–second edition (SRS-2).
ResultsThe accuracy of emotion recognition was significantly lower in individuals with ASC-ID compared to the control group with ID, especially when it came to identifying angry and fearful emotions. Participants with ASC-ID exhibited more pronounced difficulties in social functioning compared to those with ID, and there was a significant negative correlation between emotion recognition and social functioning. However, emotion recognition accounted for only 8% of the variability observed in social functioning.
ConclusionOur data indicate severe difficulties in the social-perceptual domain and in everyday social functioning in individuals with ASC-ID.
To determine the association between occupational factors, particularly psychosocial factors, and hypertension.
Descriptive and analytical cross-sectional study using logistic multivariate regression.
Fifteen cotton ginning plants in Benin.
Permanent and occasional workers in the cotton ginning industry.
Data on sociodemographic, occupational, behavioural and clinical history characteristics were collected using a number of standardised, interviewer-administered questionnaires. These questionnaires were based on the WHO’s non-communicable disease questionnaire, Karasek questionnaire and Siegrist questionnaire. Weight, height and blood pressure were measured. Any worker with systolic blood pressure ≥140 mm Hg and/or diastolic blood pressure ≥90 mm Hg according to the WHO criteria was considered hypertensive, as was any subject on antihypertensive treatment even if blood pressure was normal.
A total of 1883 workers were included, with a male to female ratio of 9.08. Of these, 510 suffered from hypertension (27.1%, 95% CI 25.1 to 29.2). In the multivariate analysis, the risk factors identified were occupational stress (adjusted OR (aOR)=3.96, 95% CI 1.28 to 12.2), age ≥25 years (aOR=2.77, 95% CI 1.55 to 4.96), body mass index of 25–30 kg/m2 (aOR=1.71, 95% CI 1.32 to 2.2), body mass index >30 kg/m2 (aOR=2.74, 95% CI 1.84 to 4.09), permanent worker status (aOR=1.66, 95% CI 1.44 to 2.41) and seniority in the textile sector >5 years (aOR=2.18, 95% CI 1.7 to 2.8). Recognition at work emerged as an effect-modifying factor subject to stratification.
Occupational factors, particularly job strain and recognition at work, are modifiable factors associated with hypertension in the ginning plants sector and deserve to be corrected through occupational health promotion and prevention.
The objective of this study was to develop clinical classifiers aiming to identify prevalent ascending aortic dilatation in patients with bicuspid aortic valve (BAV) and tricuspid aortic valve (TAV).
A prospective, single-centre and observational cohort.
The study involved 543 BAV and 491 TAV patients with aortic valve disease and/or ascending aortic dilatation, excluding those with coronary artery disease, undergoing cardiothoracic surgery at the Karolinska University Hospital (Sweden).
Predictors of high risk of ascending aortic dilatation (defined as ascending aorta with a diameter above 40 mm) were identified through the application of machine learning algorithms and classic logistic regression models.
Comprehensive multidimensional data, including valve morphology, clinical information, family history of cardiovascular diseases, prevalent diseases, demographic details, lifestyle factors, and medication.
BAV patients, with an average age of 60.4±12.4 years, showed a higher frequency of aortic dilatation (45.3%) compared with TAV patients, who had an average age of 70.4±9.1 years (28.9% dilatation, p
Cardiovascular risk profiles appear to be more predictive of aortopathy in TAV patients than in patients with BAV. This adds evidence to the fact that BAV-associated and TAV-associated aortopathy involves different pathways to aneurysm formation and highlights the need for specific aneurysm preventions in these patients. Further, our results highlight that machine learning approaches do not outperform classical prediction methods in addressing complex interactions and non-linear relations between variables.
To test the feasibility of a randomised controlled trial (RCT) of a novel preoperative tailored sleep intervention for patients undergoing total knee replacement.
Feasibility two-arm two-centre RCT using 1:1 randomisation with an embedded qualitative study.
Two National Health Service (NHS) secondary care hospitals in England and Wales.
Preoperative adult patients identified from total knee replacement waiting lists with disturbed sleep, defined as a score of 0–28 on the Sleep Condition Indicator questionnaire.
The REST intervention is a preoperative tailored sleep assessment and behavioural intervention package delivered by an Extended Scope Practitioner (ESP), with a follow-up phone call 4 weeks postintervention. All participants received usual care as provided by the participating NHS hospitals.
The primary aim was to assess the feasibility of conducting a full trial. Patient-reported outcomes were assessed at baseline, 1-week presurgery, and 3 months postsurgery. Data collected to determine feasibility included the number of eligible patients, recruitment rates and intervention adherence. Qualitative work explored the acceptability of the study processes and intervention delivery through interviews with ESPs and patients.
Screening packs were posted to 378 patients and 57 patients were randomised. Of those randomised, 20 had surgery within the study timelines. An appointment was attended by 25/28 (89%) of participants randomised to the intervention. Follow-up outcomes measures were completed by 40/57 (70%) of participants presurgery and 15/57 (26%) postsurgery. Where outcome measures were completed, data completion rates were 80% or higher for outcomes at all time points, apart from the painDETECT: 86% complete at baseline, 72% at presurgery and 67% postsurgery. Interviews indicated that most participants found the study processes and intervention acceptable.
This feasibility study has demonstrated that with some amendments to processes and design, an RCT to evaluate the clinical and cost-effectiveness of the REST intervention is feasible.
Our main objective is to assess the inter-reviewer reliability (IRR) reported in published systematic literature reviews (SLRs). Our secondary objective is to determine the expected IRR by authors of SLRs for both human and machine-assisted reviews.
We performed a review of SLRs of randomised controlled trials using the PubMed and Embase databases. Data were extracted on IRR by means of Cohen’s kappa score of abstract/title screening, full-text screening and data extraction in combination with review team size, items screened and the quality of the review was assessed with the A MeaSurement Tool to Assess systematic Reviews 2. In addition, we performed a survey of authors of SLRs on their expectations of machine learning automation and human performed IRR in SLRs.
After removal of duplicates, 836 articles were screened for abstract, and 413 were screened full text. In total, 45 eligible articles were included. The average Cohen’s kappa score reported was 0.82 (SD=0.11, n=12) for abstract screening, 0.77 (SD=0.18, n=14) for full-text screening, 0.86 (SD=0.07, n=15) for the whole screening process and 0.88 (SD=0.08, n=16) for data extraction. No association was observed between the IRR reported and review team size, items screened and quality of the SLR. The survey (n=37) showed overlapping expected Cohen’s kappa values ranging between approximately 0.6–0.9 for either human or machine learning-assisted SLRs. No trend was observed between reviewer experience and expected IRR. Authors expect a higher-than-average IRR for machine learning-assisted SLR compared with human based SLR in both screening and data extraction.
Currently, it is not common to report on IRR in the scientific literature for either human and machine learning-assisted SLRs. This mixed-methods review gives first guidance on the human IRR benchmark, which could be used as a minimal threshold for IRR in machine learning-assisted SLRs.
CRD42023386706.
While glucocorticoid (GC) treatment initiated for COVID-19 reduces mortality, it is unclear whether GC treatment prior to COVID-19 affects mortality. Long-term GC use raises infection and thromboembolic risks. We investigated if patients with oral GC use prior to COVID-19 had increased mortality overall and by selected causes.
Population-based observational cohort study.
Population-based register data in Sweden.
All patients infected with COVID-19 in Sweden from January 2020 to November 2021 (n=1 200 153).
Any prior oral GC use was defined as ≥1 GC prescription during 12 months before index. High exposure was defined as ≥2 GC prescriptions with a cumulative prednisolone dose ≥750 mg or equivalent during 6 months before index. GC users were compared with COVID-19 patients who had not received GCs within 12 months before index. We used Cox proportional hazard models and 1:2 propensity score matching to estimate HRs and 95% CIs, controlling for the same confounders in all analyses.
3378 deaths occurred in subjects with any prior GC exposure (n=48 806; 6.9%) and 14 850 among non-exposed (n=1 151 347; 1.3%). Both high (HR 1.98, 95% CI 1.87 to 2.09) and any exposure (1.58, 1.52 to 1.65) to GCs were associated with overall death. Deaths from pulmonary embolism, sepsis and COVID-19 were associated with high GC exposure and, similarly but weaker, with any exposure. High exposure to GCs was associated with increased deaths caused by stroke and myocardial infarction.
Patients on oral GC treatment prior to COVID-19 have increased mortality, particularly from pulmonary embolism, sepsis and COVID-19.
For artificial intelligence (AI) to help improve mental healthcare, the design of data-driven technologies needs to be fair, safe, and inclusive. Participatory design can play a critical role in empowering marginalised communities to take an active role in constructing research agendas and outputs. Given the unmet needs of the LGBTQI+ (Lesbian, Gay, Bisexual, Transgender, Queer and Intersex) community in mental healthcare, there is a pressing need for participatory research to include a range of diverse queer perspectives on issues of data collection and use (in routine clinical care as well as for research) as well as AI design. Here we propose a protocol for a Delphi consensus process for the development of PARticipatory Queer AI Research for Mental Health (PARQAIR-MH) practices, aimed at informing digital health practices and policy.
The development of PARQAIR-MH is comprised of four stages. In stage 1, a review of recent literature and fact-finding consultation with stakeholder organisations will be conducted to define a terms-of-reference for stage 2, the Delphi process. Our Delphi process consists of three rounds, where the first two rounds will iterate and identify items to be included in the final Delphi survey for consensus ratings. Stage 3 consists of consensus meetings to review and aggregate the Delphi survey responses, leading to stage 4 where we will produce a reusable toolkit to facilitate participatory development of future bespoke LGBTQI+–adapted data collection, harmonisation, and use for data-driven AI applications specifically in mental healthcare settings.
PARQAIR-MH aims to deliver a toolkit that will help to ensure that the specific needs of LGBTQI+ communities are accounted for in mental health applications of data-driven technologies. The study is expected to run from June 2024 through January 2025, with the final outputs delivered in mid-2025. Participants in the Delphi process will be recruited by snowball and opportunistic sampling via professional networks and social media (but not by direct approach to healthcare service users, patients, specific clinical services, or via clinicians’ caseloads). Participants will not be required to share personal narratives and experiences of healthcare or treatment for any condition. Before agreeing to participate, people will be given information about the issues considered to be in-scope for the Delphi (eg, developing best practices and methods for collecting and harmonising sensitive characteristics data; developing guidelines for data use/reuse) alongside specific risks of unintended harm from participating that can be reasonably anticipated. Outputs will be made available in open-access peer-reviewed publications, blogs, social media, and on a dedicated project website for future reuse.
To explore whether large language models (LLMs) Generated Pre-trained Transformer (GPT)-3 and ChatGPT can write clinical letters and predict management plans for common orthopaedic scenarios.
Fifteen scenarios were generated and ChatGPT and GPT-3 prompted to write clinical letters and separately generate management plans for identical scenarios with plans removed.
Letters were assessed for readability using the Readable Tool. Accuracy of letters and management plans were assessed by three independent orthopaedic surgery clinicians.
Both models generated complete letters for all scenarios after single prompting. Readability was compared using Flesch-Kincade Grade Level (ChatGPT: 8.77 (SD 0.918); GPT-3: 8.47 (SD 0.982)), Flesch Readability Ease (ChatGPT: 58.2 (SD 4.00); GPT-3: 59.3 (SD 6.98)), Simple Measure of Gobbledygook (SMOG) Index (ChatGPT: 11.6 (SD 0.755); GPT-3: 11.4 (SD 1.01)), and reach (ChatGPT: 81.2%; GPT-3: 80.3%). ChatGPT produced more accurate letters (8.7/10 (SD 0.60) vs 7.3/10 (SD 1.41), p=0.024) and management plans (7.9/10 (SD 0.63) vs 6.8/10 (SD 1.06), p
This study shows that LLMs are effective for generation of clinical letters. With little prompting, they are readable and mostly accurate. However, they are not consistent, and include inappropriate omissions or insertions. Furthermore, management plans produced by LLMs are generic but often accurate. In the future, a healthcare specific language model trained on accurate and secure data could provide an excellent tool for increasing the efficiency of clinicians through summarisation of large volumes of data into a single clinical letter.
In 2015 and 2016, WHO issued guidelines on HIV testing services (HTS) highlighting recommendations for a strategic mix of differentiated HTS approaches. The policy review examines the uptake of differentiated HTS approaches recommendations in national policies.
Data were extracted from national policies published between January 2015 and June 2019. The WHO-recommended HTS approaches included facility-based testing, community-based testing, HIV self-testing and provider-assisted referral (or assisted partner notification). Other supportive recommendations include pre-test information, post-test counselling, lay provider testing and rapid testing. Descriptive analyses were conducted to examine inclusion of recommendations in national policies.
Of 194 countries worldwide, 65 published policies were identified; 24 WHO Africa region (AFR) countries (51%, 24/47), 21 WHO European region (EUR) (40%, 21/53), 6 WHO Eastern Mediterranean region (EMR) (29%, 6/21), 5 Pan-American region (AMR) (14%, 5/35), 5 Western Pacific Region (WPR) (19%, 5/27) and 4 WHO South East Asia Region (SEAR) (36%, 4/11). Only five countries included all recommendations. 63 included a minimum of one. 85% (n=55) included facility-based testing for pregnant women, 75% (n=49) facility-based testing for key populations, 74% (n=48) community-based testing for key populations, 69% (n=45) rapid testing, 57% (n=37) post-test counselling, 45% (n=29) lay provider testing, 38% (n=25) HIV self-testing, 29% (n=19) pre-test information and 25% (n=16) provider-assisted referral. The proportion in each region that included at least one recommendation were: 100% AFR (24/47), 100% EMR (6/6), 100% AMR (5/5), 100% WPR (5/5), 100% SEAR (4/4) and 95% EUR (20/21). AFR followed by EMR included the highest number of reccomendations.
There was substantial variability in the uptake of the WHO-differentiated HTS recommendations. Those in EMR included the most WHO-differentiated HTS recommendation followed by AFR. Countries within AMR included the least number of recommendations. Ongoing advocacy and efforts are needed to support the uptake of the WHO-differentiated HTS recommendations in country policies as well as their implementation.